UPS on whole manufacturing plant.

[tkb]

I have a potential customer that wants to put a UPS and generator on their whole manufacturing plant.
Seems that their process is very sensitive to brown outs from the POCO and they lose a lot of $$$ when a brown out occurs.

They seem to be drawing about 650 amps at 480 volts and have two 150 ton chillers and many motors on their system.

They said that since there are many systems like cooling pumps and whatever else that runs their machines, it is too much trouble to separate the electrical to just what they want to back up, so they want to do the whole building.

Since there is the potential for a large inrush when a chiller starts or one of the large motors start, I don't think a static UPS will be the right application.
My Liebert guy agrees.

Does anyone have any experience with a rotary or flywheel UPS?

Also to determine the size required we want to record the service for a couple of weeks, but they insist they don't need that done since they can get a power profile from their electrical energy supplier that will provide our power utilization hour by hour for a full year.

I already told them that this is big bucks.
Once they get my price for this I think that moving the electrical circuits around to one distribution panel might be wiser so they are not backing up the the lights, AC and offices along with any other things that have nothing to do with their process.

 

[augie47]

..............I already told them that this is big bucks.
Once they get my price for this I think that moving the electrical circuits around to one distribution panel might be wiser so they are not backing up the the lights, AC and offices along with any other things that have nothing to do with their process.

I am very sure you are correct. If they are like most they will even suffer "sticker shock" at the cost of a UPS for "critical circuits" and a whole plant generator back-up.

 

[drbond24]

When they see the price for that, they'll decide they don't want it so bad anymore.

I worked at a manufacturing plant and had similar conversations for the same reasons. This plant was at the end of a transmission line, so every little thing that happened anywhere on the line effected them.

Anyway, they had a guy from Omniverter come in and give a presentation. They have voltage conditioners that can catch a brown out and compensate for it so that it doesn't effect the building. It would probably be cheaper than a UPS for the whole building. It doesn't help during blackouts, but brownouts would no longer be a problem. It would probably still be a good idea to put the critical loads in one panel.

Omniverter Active Voltage Conditioners

 

[tkb]

They also wanted me to investigate an alternate circuit from the POCO instead of the generator.
I drove up and down their street and saw only one circuit from the POCO for about 1/2 mile each way.
So that looks out of the question.
I am sure that this would cost many times more than a generator since the POCO would have to build a new pole line form who knows where and add an additional transformer.
They would have to pay for this since there would be no additional revenue for the POCO.

Oh yea, they said that after I give them my proposal they will develop a scope and spec so they can get other pricing.
Like I'm going to let that happen.:grin:

The first proposal is going to be an engineering proposal.
We should get paid for the design if we are developing the scope.

 

[Besoeker]

I have a potential customer that wants to put a UPS and generator on their whole manufacturing plant.
Seems that their process is very sensitive to brown outs from the POCO and they lose a lot of $$$ when a brown out occurs.

They seem to be drawing about 650 amps at 480 volts and have two 150 ton chillers and many motors on their system.

They said that since there are many systems like cooling pumps and whatever else that runs their machines, it is too much trouble to separate the electrical to just what they want to back up, so they want to do the whole building.

Since there is the potential for a large inrush when a chiller starts or one of the large motors start, I don't think a static UPS will be the right application.
My Liebert guy agrees.

Does anyone have any experience with a rotary or flywheel UPS?

Also to determine the size required we want to record the service for a couple of weeks, but they insist they don't need that done since they can get a power profile from their electrical energy supplier that will provide our power utilization hour by hour for a full year.

I already told them that this is big bucks.
Once they get my price for this I think that moving the electrical circuits around to one distribution panel might be wiser so they are not backing up the the lights, AC and offices along with any other things that have nothing to do with their process.

I think you are right.
Backing non-essential loads would be expensive and unneccessary.

FWIW, a few years ago I had some involvement with flywheel energy storage development project. We supplied the 1500V 600A converter to run them up and a similarly rated braking system to load them.
For us, it was a contract to design, supply and commission. For the organisation undertaking the development, it was a funded project. But it died a death. I don't know all the details but I think funding ran out or was withdrawn.
Either way, the system didn't make it into the commercial world.

 

[dbuckley]

In the world of data centres, rotary UPS means just one thing: Piller

The combined UPS/Generator - see here - would be just the ticket. Note that these units can be supplied with oversized diesel and motor/generator components, so you could have true no-break power to most of the plant, and seperate the chillers off to short break power (20 seconds outage) and thus save a load of money by downsizing the actual UPS load requirement. Talk it through with the Piller folks they'll give you some options.

These things are not cheap, but they are very impressive pieces of equipment.

They are also a load less fussy about their environment than semiconductor UPSs.

Edited to note: Backing up the whole building isn't a completely mad idea. If you're a heavy plant then the lighting and offices probably only consume a few percent of the overall load, and thus have a 2000KVA UPS rather than a 1800KVA isn't a really big deal. Plus you get to keep the office workers going. Remember, they're the people who actually deposit the salaries into the bank, so its really important that Payroll keep power

 

[LarryFine]

Oh yea, they said that after I give them my proposal they will develop a scope and spec so they can get other pricing.
Like I'm going to let that happen.:grin:

The first proposal is going to be an engineering proposal.
We should get paid for the design if we are developing the scope.

Absolutely. Anything they can use, whether you do the job or not, you should receive compensation for.

 

[hurk27]

Talk with some of these company's that manufacture the UPS's as many will come out and do all the leg work for you just to get the sale. this can save allot of headaches down the road also when something doesn't work out, as it's there baby.

 

[weressl]

I have a potential customer that wants to put a UPS and generator on their whole manufacturing plant.
Seems that their process is very sensitive to brown outs from the POCO and they lose a lot of $$$ when a brown out occurs.

They seem to be drawing about 650 amps at 480 volts and have two 150 ton chillers and many motors on their system.

They said that since there are many systems like cooling pumps and whatever else that runs their machines, it is too much trouble to separate the electrical to just what they want to back up, so they want to do the whole building.

Since there is the potential for a large inrush when a chiller starts or one of the large motors start, I don't think a static UPS will be the right application.
My Liebert guy agrees.

Does anyone have any experience with a rotary or flywheel UPS?

Also to determine the size required we want to record the service for a couple of weeks, but they insist they don't need that done since they can get a power profile from their electrical energy supplier that will provide our power utilization hour by hour for a full year.

I already told them that this is big bucks.
Once they get my price for this I think that moving the electrical circuits around to one distribution panel might be wiser so they are not backing up the the lights, AC and offices along with any other things that have nothing to do with their process.

Generator for the entire load, including the UPS, and UPS only for the critical control type loads such as PLC's and DCS's and autonomous control systems on packaged equipment. It should be an engineered system.

 

[tkb]

We will be going with the Piller system.
We are waiting for the load analysis before we can engineer this system.
Thanks for the suggestions.

 

[weressl]

We will be going with the Piller system.

I wouldn't.

 

[tkb]

I wouldn't.

Why? :-?

 

[charlietuna]

They would be better off contacting the US Navy for possibly ordering the type of system used on nuke subs where large motor(AC-DC) generators automatically shift from motor to generator when ac power is lost without noticable loss of frequency. The power is then provided by large batteries that are normally fully charged and floating prior to the power loss. A 650 amp-480 volt UPS is larger than i can imagine...

 

[weressl]

They would be better off contacting the US Navy for possibly ordering the type of system used on nuke subs where large motor(AC-DC) generators automatically shift from motor to generator when ac power is lost without noticable loss of frequency. The power is then provided by large batteries that are normally fully charged and floating prior to the power loss. A 650 amp-480 volt UPS is larger than i can imagine...

It's only 600kVA! Not uncommon in large data facilities. The largest one I installed was a 350kVA with a 10 min. battery bank and a 650kVA diesel backup. If the generator did not kick in in 1 min. they had 9 minutes for orderly shutdown and data salvaging. If they had no generator, they had no cooling either so the mainframes would overheat and shut down anyway.

 

[weressl]

Why? :-?

Not only do you have the main components of a static UPS but you have a motor generator in series with it. That does decrease your reliability based on simple mathematical model. Rotating and wearing components bearings as a single point of failure always make me nervous. They take up more space. They have a higher short circuit delivery; more damage during faults. They brag about galvanic isolation yet most true UPS's also have it. In the standard model they do not ofer a staic bypass, so an internal failure send you bye-bye. If you have one then it doubles the electronic switches as they have one on the input, so it blocks the inverter back to the utility. Independent bypass synchronization is an issue, it has to slow down or speed up a mass before it can switch....and so on and so forth.

 

[dbuckley]

Based on my previous runins with Piller rotary UPSs albeit a few years back: There are a number of things I either know are wrong or I think are doubtful in what weressl states, thus on this topic I'd advise caution in taking his advice, beyond the term "use due diligence" which one should always do.

For example; I dont know where the galvanic isolation claim comes from: there is no isolation between mains input and no-break output, they are the same thing because there are no active (ie electronic) components in the circuit path: Its all solid copper.

They take up more space. They have a higher short circuit delivery; more damage during faults.

They are bigger, thats true, but maybe not if you add up all the space requirements of the stuff they replace, not to mention they can live in a brick walled shed, not expecting a pretty CRAC equipped facility to live in like large semiconductor UPSs usually do.

The higher short circuit capability I rate as a definite plus; I've mentioned many times that (double conversion) semiconductor UPSs are very poor at delivering sufficient fault current for sufficient time to trip a downstream breaker, and thus these UPSs generally adopt a strategy of going to bypass on overload so the utility provides the current necessary to open the downstream breaker. 99% of the time this is fine, but if a fault occurs whilst there is no utility power then the UPS has to either produce enough current for long enough to open the breaker, which it cant, or it goes offline to prevent damage, and then you're out of power.

And also, the question of "more damage" is relative; the reason you're installing a no-break supply in the first place is that outages are expensive, thus some damage to electrical stuff whilst almost all the load remains up may be vastly preferable to no damage but the plant powered off. For many sorts of plant, failure of power means a multi-month downtime for cleanup.

It is also true that mechanical items are subject to wear and need mechanical upkeep but the absence of relatively fragile electronics is the tradeoff. The failure modes are different.

The ability to provide power reserve is also handy for starting motors.

For me and for many others in datacentres, Piller's rotary UPSs would be the first choice, if budget (and prejudice, often born of ignorance) didn't get in the way. With pairs of them in a online maintainable situation (what is now called 'Tier IV' but back in the day what I called "24 x 7 x 20 years" which is what a colo facility needs, but almost none provide) they are pretty much unbeatable as the maintenance risk window diminishes to zero.

To repeat the advice: talk to the Piller folks, take up the references they offer.

The bit that weressl quotes that is absolutely true is:

If the generator did not kick in in 1 min. they had 9 minutes for orderly shutdown and data salvaging. If they had no generator, they had no cooling either so the mainframes would overheat and shut down anyway

This is what got me into high availability power. When I was young and wet behind the ears and just started in IT I worked for a company with several large mainframes, and that had UPS backed by several 2MVA gensets and three 900KVA cogen plant all in their recently completed building.

So a roading contractor put the backhoe thru the underground 11KV taking out the building (and killing the backhoe op who had been told by his foreman the power was off). The diesels and cogens all burst into life, but they refused to syncronise. So they never buss up, so the cooling water pumps never get power, so the gensets overheat and stop. I can still recall a systems programmer watching the temperature guage on the production mainframe, waiting for it to redline and thus the mainframe to stop 'cos there was no cooling, whilst application guys rushed to shut down applications so they would close clean and not trash some part of the days work.

All those gensets potentially several times the needed capacity, but still the mainframe crashed. I thought that was stupid, and stupidity inflicted upon us by electrical engineers, who knew all about electricity but nothing about what is important in this IT age, and sold us down the river. I have fought to stop that sort of stupidity from happening wherever I can since then.

Guess thats the rant over.

 

[hurk27]

I'd advise caution in taking his advice,

I think we can state our opinions without referring to opinions of others

Either wrong or not, we all have the right to voice an opinion.

 

[weressl]

I think we can state our opinions without referring to opinions of others

Either wrong or not, we all have the right to voice an opinion.

Anybody who takes advice without due diligence, be it from me or anybody else, is a moron.:smile:

There is no need to point out that somebody is 'wrong', before stating your opinion, unless your opinion is on shaky ground. Logic, facts and data usually speak for themselves. Unless you are a 'consultant'. (Those who can't do: teach or advise.)

Trashing electrical engineers is a favored past-time here and highly amusing given the appearent ignorance of posters in engineering matters or even basic science. So keep them coming, I am amused.....:grin: I am thinking about collecting them to publish it in a book at one point.

 

[drbond24]

I thought that was stupid, and stupidity inflicted upon us by electrical engineers, who knew all about electricity but nothing about what is important in this IT age, and sold us down the river. I have fought to stop that sort of stupidity from happening wherever I can since then.

I missed the part of the situation you described that was the fault of electrical engineers...? I'm also curious what kind of consultant on the subject being discussed isn't an electrical engineer (probably with a P.E.).

 

[dbuckley]

I was a bit hasty with my keyboard with my earlier posting, and although I stand by what I state, I was out of line with an ad hominen attack for which I apologise.